1

Paper Presented at 2003 PRRES Conference, Brisbane, Australia Macro-economic Factors Affecting Office Rental Values in Southeast

Asian Cities: The case of Singapore, Hong Kong, Taipei, Kuala Lumpur and Bangkok

Wei CHIN (Henry) BA, MSc

Department of Real Estate Management

Oxford Brookes University

Headington, Oxford

OX3 0BP, UK

wchin@brookes.ac.uk Abstract: The purpose of this paper is to identify the macro-economic drivers of office rental values in

South-east Asian cities over the period 1988 � 2001 by adopting the demand and supply

framework. The paper tries to use the existing single-equation method to examine influences

on office rents in five South-east Asian cities (Singapore, Hong Kong, Taipei, Kuala Lumpur and

Bangkok). Real GDP, unemployment rate, floor space for office buildings, interest rate, lending

rate, consumer index and service sector output will be included in examining the movements of

office rental values. In spite of data limitations, this research demonstrates that in Singapore

and Taipei office markets, rental values are mainly determined by changes in office floor space,

while in Hong Kong and Kuala Lumpur office markets, lending rates have a greater effect on

rental values. However, there are no significant factors in the Bangkok office market. The

results also suggest that office markets in South-east Asia are highly linked with development

markets, while demand-side variables do not have much impact on those markets.

This paper also reviews previous empirical studies and assesses future research directions for

South-east Asian office markets.

Keywords: South-east Asia Office Rents, Single-Equation Model, Regression.

Introduction: The World Bank noted that from the 1960s to 1990s, twenty-three economies of East Asia grew

faster than all other regions of the world. Most of the growth could be attributed to eight

economies: Japan, the �Four Tiger Economies� (Singapore, Hong Kong, Taiwan, and South

Korea), as well as Indonesia, Thailand and Malaysia which were known as �Newly

2

Industrialising Economies� (NIEs). The latter started to achieve high economic growth in the

early 1990s. Nevertheless, the former have experienced high economic growth since early

1980s. These economies all achieved high economic growth and attracted large amounts of

foreign investment.

These countries have several common characteristics in their economic performance and

follow similar growth patterns (Armitage 1996, Yong 2000, Asian Development Bank 2001):

• Rapid urbanisation in emerging economies

• Sustained economic growth

• Extreme population pressure

• Accelerating rate of change in social and economic structures

• Increasing share of international economic activity

• Growing affluence

• Large domestic markets

• Social and political tension

• Swift adoption of new technology and work practices

• High technology exports accounting for large portion of GDP growth

• Higher initial levels and growth rates of human capital

• High rates of productivity growth

The relative attractiveness of a large number of alternative South-east Asian cities (such as

Hong Kong, Singapore, Taipei, Kuala Lumpur, and Bangkok) is likely to become a more

frequent consideration in the corporate investment strategies of occupiers and investors. This

research aims at establishing the relationship between macro economic activities and office

market performance in South-east Asian cities. There is much interest in South-east cities from

global multi-national property companies and property investors. However, there is a lack of

relatively detailed South-east Asian property research and data. This paper provides an

overview of the office market analysis in South-east Asian cities and methodological issues

concerning the application of Ordinary Least Square (OLS) model of office rental levels within

five major South-east Asian cities.

Previous studies have attempted to employ demand and supply variables to explain office

rental movements. There are two difficulties in this type of study. Firstly, the demand and supply

variables are proxy variables for the property markets, because there are no direct-measured

variables within the property markets. Secondly, the availability and reliability of the data are in

question (Gardiner et al. 1988, Giussani et al. 1992, Tsolocas et al. 1993, D�Arcy et al. 1997).

3

This paper follows the author�s paper to the PRRES conference in Christchurch (Chin, 2002). It

attempts to model office markets in South-east Asian cities. The paper gives a brief overview of

office markets in South-east Asian cities, previous empirical studies and reports results of an

empirical investigation of the relationship between office rents and fluctuations in economic

activities.

Overview of Southeast Asian Office Markets: Since the 1960s, economic growth in East Asian countries has been higher than in other

regions of the world (World Bank, 2000). From 1985 to 1997, the best economic performance in

global terms was in the Asia-Pacific region, with the property sector and stock market

particularly strong during that period. As a consequence, foreign investment flowed into this

area in search of huge profits. However, some countries in the region experienced similar

problems, while the accompanying risks were high. Lending to property and over-supply proved

to be key causal factors in the 1997 financial crisis throughout the region. (D�Arcy, 1998).

From the 1980s a property boom occurred in most of the Asia-Pacific property markets, which

encouraged property companies to continue building. In time, an over-supply emerged in many

cities across the region, which led ultimately to a bad debt problem in the property sector,

mainly as a result of the shallow banking system. At one time, over 200 million square feet of

space was under construction, with a large proportion in Kuala Lumpur, Bangkok and Jakarta,

and at the same time there was an increasing vacancy rate (Knight Frank 1998). It is worth

noting that rents were very high for certain classes of space, even by international standards. In

the area as a whole, differences in the property process, political uncertainty, and the regulation

of the financial markets were all important factors, and were significant in the performance of

the property market (Chin, 1999).

Investment return in property markets experienced high growth in the 1990s, attracting much

local and foreign capital. In some countries, an oversupply problem had emerged by 1996.

However, little attention was paid to this, which contributed to regional financial turmoil in 1997.

During the financial crisis, property companies experienced severe difficulties in repaying their

debts as demand for property in most sectors dropped dramatically, especially in Thailand. New

developments had to be halted when rapidly falling currencies doubled or even tripled debt

repayments, and prices and rental values falling by 30% to 40% made the difficulties even

4

worse. In Hong Kong and Malaysia, property market prices and stock markets dropped by a

large amount, and lending rates increased in these countries, leading to further falls in property

prices and values, while vacancy rates rose. Singapore and Taiwan with their strong currencies

and large reserves were able to limit this financial damage in 1997 and 1998 (Chin,1999).

South-east Asia experienced a strong economic recovery in 1999, marked by strong growth

and a robust performance after the dramatic slowdown of 1998. Consistently strong global

demand and recovering domestic demand proved to be major factors stimulating economic

activity across the region. Capital moved from old business sectors into the new economy

sectors such as the Internet and telecommunications. The obvious evidence is in the stock

market. Many of the South-east Asian stock markets were then dominated by the new

economic sectors, and these became one of the major tenants for the office market, especially

in Hong Kong, Singapore and Taipei. However, the service sector is still a major contributor to

the growth of real GDP (Asian Development Bank 2001) especially in Hong Kong, Taiwan,

Singapore, and Thailand. Table (1) shows the sectoral share of GDP in 1980, 1990 and 2000.

This can provide an idea of changes in economic structure in those countries.

Table (1): Sectoral Share of GDP (percent)

Agriculture Industry Service Country 1980 1990 2000 1980 1990 2000 1980 1990 2000 Singapore 1.3 0.4 0.1 38.1 34.4 34.3 60.6 65.3 65.6 Hong Kong 0.8 0.3 0.1 31.7 25.3 14.6 67.5 74.5 85.3 Taiwan 7.7 4.2 2.1 45.7 41.2 32.4 60.6 65.3 65.6 Malaysia NA 15.2 8.6 NA 42.2 51.7 NA 42.6 39.7 Thailand 23.2 12.5 9.1 28.7 37.2 41.7 48.1 50.3 49.2 Source: Asian Development Bank (2001), Key Indicators of Developing Asian and Pacific Countries, page 50.

An analysis of the office rental movement in those five cities from 1988 to 2001 (those figures

can be seen in the appendix), office rents achieved their peak in the first half of 1990s, owing to

the economic expansion across the region. Foreign investment flowed into the region from the

late 1980s, due to attractive rates of return and the low cost of borrowing. Office investment

markets played an important role in the anticipated high rate of economic growth. Yet

oversupply problems existed in many of the cities, such as Bangkok and Kuala Lumpur, from

the mid-1990s. Owing to oversupply, developers were unable to pay off their debts. Then the

property and stock markets crashed in 1997, and turmoil spread across the region. Office rental

levels dropped in late 1997 and early 1998. However, prime office buildings were not severely

affected.

There are some common factors existing in the five cities: the markets grew dramatically from

the late 1980s till 1997; foreign investment accounted for a huge proportion of property

5

investment; and governments supported development schemes in order to attract more

investment. However, political stability played an important role in the region, especially in the

Hong Kong and Taipei markets. These markets depend significantly on their political

relationship with China. Within all five cities, IT, as well as the finance and telecommunications

sectors, are the most active players in the markets. These sectors clearly affect growth and any

recovery in the region. In 2001, the global economic situation slowed down, especially in the

USA and Japan. The USA is one of the major trading partners for the region. Therefore, the

economic situation did not seem to be as positive as in the first half of 2001, especially after the

September 11th attack. Taiwan and Singapore suffered the lowest GDP growth and the highest

unemployment rates in a decade. Thailand, Malaysia and Hong Kong also showed signs of

slowing economic growth in the second half of 2001. Singapore, Taipei and Hong Kong were in

the market recession period; while, Bangkok and Kuala Lumpur were in the market recovery

process, according to Cushman and Wakefield Research (2001).

Previous Empirical Office Markets Reviews: Previous studies have attempted to explain or forecast office rental values by using

econometric models. Existing quantitative research on office markets has produced a variety of

econometric models, which have estimated rental movement in the different geographical

areas. Most of the research has been done in the USA, while in Europe most of the research

has been performed in the UK. There are no relevant published research papers for South-east

Asian countries.

In previous research, there were two approaches to analysis of the markets: the multi-equation

model (Kelly 1983, Rosen 1984, Hekman 1985, Shilling et al. 1987, Wheaton 1987, McClure

1991, Hendershott et al. 1997, Parker et al. 2001) and the single-equation model (Gardiner et al.

1988 1991, Giussani et al. 1993, Tsolocas et al. 1993, D�Arcy et al. 1994, D�Arcy et al. 1997,

Keogh et al. 1997, D�Arcy et al. 1998) In the former, the dependent variable in one equation

might appear as the independent variable in other equations. This approach is very popular in

American literature. In the latter, there is a single dependent variable and all the explanatory

variables are in one equation. This approach has been widely used in the British literature.

In multi-equation modelling studies, equations represent demand, supply and rents, linking with

a number of endogenous and exogenous variables. The exogenous variables normally include

employment measures, construction costs, interest rates and tax rates. The endogenous

variables are development, rents, absorption, vacancy, floor space, vacancy rates. These

6

models are widely used by American researchers, and cover the development markets and

some user markets.

For these American researchers a key point is the assumption of a natural vacancy rate. This

assumption is similar to those of the natural rate of employment in labour market studies. The

rationale for the existence of a natural vacancy rate is that property developers and property

owners, aware of the binding nature and the length of office leases, "withhold vacancy office

space in inventory to capitalise on opportunities to supply at higher rental during periods of

increased demand" (Shilling et al. 1987 page 91). Deviations away from the natural vacancy

rate have been used as a measure of the short-term demand and supply conditions in the office

market and as a determinant of rents. If demand for space is high and the actual level of office

vacancies is below the natural vacancy rate, upward pressure will be exerted on rents. However,

if the number of vacancies is higher than the natural vacancy rate, rents will face downward

pressure. Therefore, rents move more rapidly the further the actual vacancy rate moves away

from the natural vacancy rate in either direction. Rosen (1984). Hekman (1985), Shilling et al.

(1987), and Wheaton and Torto (1988) use this idea to research aspects of the office user

market and development market. In the UK, Hendershott et al. (1997) use the same idea for

examining London office markets.

Multi-equation modelling is more theorised than single-equation modelling. However, there are

no consistent forms in this area. If certain functions of the markets are not described adequately

by equations, the result can be spurious (D�Arcy et al. 1998, Wong 2002).

The existing literature using multi-equation modelling has certain characteristics (Kelly 1983,

Rosen 1984, Hekman 1985, Shilling et al. 1987, Wheaton 1987, McClure 1991, Pallakowski et

al. 1992, Hendershott et al. 1997, MacFariane et al. 1999, Parker et al. 2001):

• These models usually comprise two to three equations which include demand,

development and rental change. The equations consist of some endogenous variables,

such as absorption, rent, completed development, vacancy rates, and floor space; and also

a number of exogenous variables, such as the employment rate, interest rate, rate of

finance, construction costs and inflation rate.

• Rental changes are always modelled by the vacancy rates with some various gaps, and the

variable is always significant in the model.

• The development (supply side) model is mainly driven by the profitability formulation.

Vacancy rate is always significant in the model.

• Some of the American studies (Kelly 1983, Wheaton et al. 1983) suggest that measures of

economic activity are not all equally effective in explaining office rental values (Giussani et

7

al. 1993a), but disaggregated service sector employment variables appear to be more

consistently significant.

In their results, the vacancy rate is always significant among those studies. However, the use of

a natural vacancy rate is another issue to be confronted. Most of the existing literature uses

average vacancy rate as the natural vacancy rate, which is doubtful in theory, because of the

way of calculating the natural vacancy rate.

The single-equation method generally tries to catch the relationship between economic activity

and office rental markets by using a demand and supply framework. The theory justifying this

method is based on the interaction of demand and supply. Rental values in the short term are

the result of disequilibrium situations in real property markets. In the majority of empirical work

on rent determination, the model used takes the form of an equation where both demand and

supply determinants are included (Giussani et al. 1993a, Giussani et al. 1993b, Gardiner et al.

1988, 1991; Tsolocas et al. 1993, D�Arcy et al. 1994, Newell et al, 1996, D�Arcy et al. 1997,

Keogh et al. 1997, Tsolocas et al. 1998, D�Arcy 1998).

Previous empirical studies in this area have examined a wide selection of demand and supply

proxy variables in an attempt to predict office market performance. Office rental values can be

modelled by using the theoretical demand-supply framework, which has been successfully

employed by using a wide range of variables to proxy demand and supply influences, even if

the availability and reliability of the data can be questioned. Various studies exist for single

countries, particularly for European cities or sectoral office market performance, generally

focussing on a set of common variables to determinate rental values.

Since the mid-1980�s, commercial property forecasting services have relied on this technique,

given the historically stable relationship between real rent levels and demand-side and

supply-side variables. Commercial property performance can be predicted by the movement of

activity of the economic sectors. Previous research in Australia and in the UK has focussed on

linking commercial property with a variety of underlying economic factors (Giussani et al. 1993a,

Giussani et al. 1993b, Hedershott 1995, Higgins 1996, Higgins et al. 2000, RICS 1994, Keogh

et al. 1995, McGough et al. 1995 1997, Tslolacos 1998, D�Arcy 1999). In searching through

previous literature about South-east Asian cities, not a single relevant research paper was

found; the key literature review will therefore focus on relevant European papers.

Gardiner and Henneberry (1988, 1991) attempt to research rent determination in the standard

planning regions, and do so using spatially disaggregated annual data for the period 1977 to

1984. They find that regional GDP and the regional stock of floor space are the main factors

8

effecting office rents in expanding regions. However, previous rental levels and GDP influence

current rental levels in declining regions.

Giussani, Hsia and Tsolocas (1993) are the first to attempt to examine rent determinants across

European cities, and their work can be used as a framework to examine Southeast Asian cities.

They examine the relationship between rental value and economic activity by using

cross-section and time series analysis (OLS methods). Again, this research is based on a

demand and supply framework. However, the study ignores the supply-side variable because

data is not available across Europe. They assert that the demand-side variable can explain

office rental markets across Europe very well. In spite of limited data, the results show that GDP

and unemployment rates play an important part in determining office rents. However, if

supply-side variables had been available and included, the model would have been more

theoretically robust.

After the comparative research by Giussani et al. (1993), D�Arcy, McGough and Tsolocas (1994)

examine the determinants of office rental values in twelve European cities over the period

1982-1993. Their result is consistent with the results obtained by Giussani et al. (1993). They

conclude that GDP and unemployment rates are the most important common determinants of

changes in rental value across those twelve cities. They also suggest that larger markets (in

terms of total office stocks) are better modelled by the standard explanatory variables; this

appears consistent with the American study (Pollakowski et al. 1992).

After finding that market size might affect determinants of office rental values, D�Arcy, McGough

and Tsolocas (1997) examine the relationship between office rents, national economic

conditions, market size and the measure of economic growth and changes in 22 European

cities� economies over the period 1982 � 1994 using time series cross-section methodology.

The result shows that local factors still seem less influential on European cities relative to

national factors. They again miss out the supply-side data, because of the lack of availability

and poor quality of such data.

Keogh, McGough and Tsolocas (1998) model the user, investment, and development elements

of British office markets. Theirs is the first research paper which examines three sub-markets:

user, investment and development. They use econometric models to estimate rents in user

markets, capital value for investment markets and new volume of new office space for

development markets. Their research is the first empirical investigation of the demand and

supply framework in all three sub markets. Their empirical result shows that the investment

market model does not produce a satisfactory result. The user market has a better result which

is the same as in the study of RICS (1994).

9

In other single-equation modelling studies, Dobson and Goddard (1992) provide an insight into

the determination of office rents in certain regions across the UK. In their research, only

demand-side variables have been used to test the determination of rental levels, which is weak

in theory, and they do not make a clear relationship between owner-occupied and rental

markets. McGough and Tsolocas (1994, 1995) use vector autoregressive (VAR) and ARIMA

methods to predict short-term office rental value forecast in the UK. They demostrate the

importance of the service sector and employment in banking, insurance and finance in office

rent modelling. Other research papers examine individual markets and offer additional insights

into factors affecting office rental movements (McGough et al. 1998, D�Arcy et al 1998). Both

papers find that GDP and employment rates are important factors. Also there are some lagged

effects on office space affecting rental movements. (Keogh et al. 1994, McGough et al. 1998,

D�Arcy et al 1998). There are also a few other studies employing single-equation modelling in

Australia and USA, such as Brennan et al. (1984), Newell et al. (1996). They also use macro,

financial and spatial variables to explain movements in the property markets.

From this review, it can be seen that there are no relevant empirical studies for South-east

Asian cities or countries. The author seeks to determine the main macroeconomic factors

influencing office rental values in five South-east Asian cities.

The use of either the multi-equation model (structure model) or the single-equation (reduced

model) model has to face the problem of availability and reliability of data. The usage of the

respective models has to be based on data availability and quality in order to provide the best

result. In many markets, significant problems exist in relation to the availability of suitable

indicators of supply, which results in there being only demand-side influences in those models.

In general, regression models based on key macroeconomic variables have provided good

results over limited periods of time, but have tended to break down over longer periods (Lee

1998).

There are some typical demand side variables for a model of office rent determination, which

include GDP/unemployment rate/service sectors output/employment and interest rates as

demand-side variables. However, floor space is usually used as a supply-side variable.

• Real GDP

The variables affecting the demand-side are most likely to be economics-based. These

variables are normally approximated by real GDP and employment rates in the service sector

(D�Arcy et al. 1998). GDP normally represents general economic conditions, as does the

10

unemployment rate. In single-equation research, the real GDP figure remains a consistently

significant influence on office rental markets. Barras (1983) noted that real GDP is the most

appropriate and widely used demand-side measurement at an aggregate level, which gives a

broad indicator of office activity, including both for manufacturing and service sectors of the

economy. This assumption gains a lot of support from empirical studies (Gardiner et al. 1988,

Dobson et al. 1992, Giussani et al. 1993, RICS 1994, D�Arcy et al. 1994, McGough et al. 1995,

Tsolocas et al. 1998. D�Arcy et al. 1998, McGough et al. 1998, Higgins et al. 2000, MacFarlane

et al. 2000).

• Employment/unemployment rates in the service sector

Employment/unemployment rates normally proxy for economic conditions, and the service

sector employment rate has close links with office rental values according to previous studies

(Gardiner et al. 1988, Dobson et al. 1992, Giussani et al. 1993a, Giussani et al. 1993b, D�Arcy

et al. 1994, McGough et al. 1994, Hendershott et al. 1996 Keogh et al. 1998, Tsolocas et al.

1998, D�Arcy et al. 1998). It is believed that the majority of service sector activity takes place in

an office environment (Dobson et al. 1992, D�Arcy et al. 1994, 1997, Keogh et al. 1998).

Demand for office space is associated with employment level changes. It is a derived demand,

reflecting the demand for the services produced by office-based activities. Changes in the level

of employment in the service sector are expected to reflect the trends in particular office-based

industries which have generally been considered to be the most dynamic (Giussani et al. 1993a,

Giussani et al. 1993b, McGough et al. 1998). When employment in the service sector rises, it

increases demand for office space and pressure rises on the rental values.

• Service sector output

Service sector output is a similar measure to service sector employment or unemployment. It is

used as an alternative measure of national activity and performance in service sector industries,

as the main occupier of office buildings is the service sector (D�Arcy et al. 1995, Keogh et al.

1998).

• Interest rates:

Interest rates provide an indication of the availability and cost of capital, and also they are

considered as predictors of economic conditions. (Chang et al. 1999). There are a few articles

using interest rates as a variable to examine office rental movements. (Giussani et al. 1993,

D�Arcy et al. 1994, Keogh et al. 1998). High interest rates will discourage development

decisions, and then rental values will increase. However, none of these empirical studies shows

that interest rates have a significant effect on rental value. Interest rates can also indicate

directions of monetary policy and the dampening effect of high interest rates on economic

activity and well established.

11

• Office floor space

In some studies, the supply-side variables are omitted, because of data non-availability. There

are a number of supply-side variables in the existing literature used to investigate influence on

office rents. They mainly focus on office floor space, and such issues as changes in total stock

of office space (Gardiner et al. 1988); the volume of new office construction (Keogh et al. 1998,

D�Arcy et al. 1998); the level of new orders for office spaces (Giussani and Tsolocas 1993a); the

completion of office and retail space (Tsolocas et al. 1998); and changes in the volume of office

building output (McGough et al. 1994). Although some of the measurements are not perfectly

represented by supply-side variables, the use of a more appropriate variable is restricted by

data non-availability. The empirical results demonstrate that office floor space does have some

significant effects on office rental markets, but find it is less influential than demand side

variables (Hekman 1985, Gardiner et al. 1988, RICS 1994, Keogh et al. 1998, D�Arcy et al.

1998).

More than 20 variables have been tested as explanatory variables in the existing literature over

the past 20 years. Table 2 shows the variables matrix which has been used in existing literature.

This variables matrix shows that the majority of the single-equation models attempt to use

macro factors and spatial factors (as supply-side variables) to examine office rental movements.

Multi-equation models are more likely to use spatial factors and financial factors to explain the

office rental adjustment process.

Table 2: Variables Matrix

12

Explanatory Variables /

Multi-Equation

Single EquationGDP GNP

Employment Rate or

Unemployment

rate(Total/Service)

Sectors)

Interest

rates (long

term or

short term)

Inflation Economic

UncertaintyIncome Population

Authors Year MACRO FACTORS

Kelly 1983 M *Rosen 1984 M * * Hekman 1985 M * * Shilling, Sirmans, Corgel 1987 M * * Wheaton 1987 M * * Gardiner and Henneberry 1988 S * * * Wheaton and Torto 1988 M McClure 1991 M * * * Gardiner and Henneberry 1991 S Dobson and Goddard 1992 S * * Pollakowski, Wachter, Lynford 1992 M * Giussani, Hsia, Tsolocas 1993a S * * * * D'Arcy, McGough, Tsolocas 1994 S * * * * McGough and Tsolocas 1994 S * * * * McGough and Tsolocas 1995 S Hendershott, Lizieri, Matysiak 1996 M * * * D'Arcy, McGough, Tsolocas 1997 S * * Wheaton, Torto, Evans 1997 M * * Keogh, McGough, Tsolocas 1998 S * * * D'Arcy, McGough, Tsolocas 1998 S * * McGough, Olkkonen, Tsolocas 1998 S * * MacFarland and Moon 2000 M * Parker, MacFariane, Whiley 2001 M * * *

Continue:

13

Explanatory Variables

Office-floor Space

(Total/New/ Changes)

Vacancy Absorption Past

Rental Value

House Index

Share Price

Bond and Tbill Yields

Construction Cost

Cost of Capital

Operating Expenses

Yield/ Capital value

Tax

Authors Year SPATIAL FACTORS FINANCIAL FACTORS

Kelly 1983 * * * Rosen 1984 * * * * * *Hekman 1985 * * * * Shilling, Sirmans, Corgel 1987 * * * Wheaton 1987 * * * * * Gardiner and Henneberry 1988 * Wheaton and Torto 1988 * * McClure 1991 * * * Gardiner and Henneberry 1991 * Dobson and Goddard 1992 * Pollakowski, Wachter, Lynford 1992 * * * * * Giussani, Hsia, Tsolocos 1993a D'Arcy, McGough, Tsolocas 1994 * * McGough and Tsolocas 1994 * * * McGough and Tsolocas 1995 * Hendershott, Lizieri, Matysiak 1996 * * * * * * * D'Arcy, McGough, Tsolocas 1997 * Wheaton, Torto, Evans 1997 * * * * * * Keogh, McGough, Tsolocas 1998 * * * D'Arcy, McGough, Tsolocas 1998 * McGough, Olkkonen, Tsolocas 1998 * MacFarland and Moon 2000 * * * Parker, MacFariane, Whiley 2001 * * *

Methodology:

The theoretical part of this research is constructed by use of the demand and supply

relationship. When property markets are in equilibrium, changes in demand will tend to

generate new supply. Before new supply reaches the markets, the price will increase. However,

when markets are in disequilibrium, demand and supply relationship will be unbalanced, which

will potentially result in over supply or shortage of property in the market. This will cause

fluctuations in the price (Fisher 1992, Giussani et al. 1993a, 1993b, Keogh 1994, Morrison 1994,

Tsolocas et al. 1998, D�Arcy et al. 1998). Office rental values in the short term exhibit similar

results when there is imbalance in the office market.

14

In this study, the single-equation model of office rent determination is adopted. Office rental

value can be determined by the interaction of demand and supply factors which affect office

rental markets. In this model, office rental values in each South-east Asian city are determined

by demand side factors (Demand i) and supply side factors (Supply i) affecting its office market

in period t. Thus

Rentit= (Demandjit, Supply lit) (1)

Where Rent it is the net office rent in city i in period t and j is the number of demand side

variables and I is the number of supply side variables

Assuming that the relationships between variables are linear, the rental level can be expressed

in the following terms:

Rentit= A +B* Demandjit+ C * Supply Iit+ e (2)

where Demand i represents demand-side variables;

Supply represents supply-side variables and e is an error term. Equation (1) and (2) provide the

general model of the single-equation for South-east Asian cities empirical investigation. This

model is similar to that constructed by Giussani et al 1993a, 1993b, D�Arcy et al. 1994, Morrison

1994, Keogh et al. 1998, D�Arcy et al. 1998

The models of office rents include both demand-side variables and supply-side variables. In the

author�s model, net office rents are used. Real GDP (GDP), the rate of unemployment (U),

interests rates (IR), lending rates (LR), consumer index (CI) and service sector output (Ser) are

the proxy variables of demand-side. Changes in floor space (FS) are the proxy variables of

supply-side.

Substituting the above variables into equation (2), it emerges as

Rents it = A + B * GDP it + C * U it + D * IR it + E * LR it + F *CI it + G * Ser it + H * FS it + e

(3)

Where i represents different cities and t represents different periods.

There are two problems associated with this type of approach. Firstly, the time series data for

South-east Asian cities are limited to 14 annual observations from 1988 to 2001. On such

limited data, equation (3) will be the result of the lack of degree of a freedom. Therefore, results

will be affected. Secondly, there are also likely to be close relationships between dependent and

15

independent variables in equation (3). This suggests that there might be a multicollinearity

problem existing in equation (3) (Giussani et al. 1993a, 1993b D�Arcy et al. 1994, Morrison

1994). In order to resolve this problem, it is reasonable to estimate the effect of those

dependant variables individually on real office rental value. The equation can be explained as

Rent it = A + B * Variables jit + e (4)

Equation (4) will allow us to determine the variables which explain South-east Asian cities;

office rental movements. However, because of limitations in time-series data, those data are

marked by a trend, the equation (4) result might cause a spurious correlation. In order to

eliminate this effect, the first difference method will be applied to all variables. Therefore, the

equation can be re-written as

∆1 Rent it = A +∆1 Variables jit + e (5)

where ∆1 is the first difference operator. After the variable that explains the largest variation in

growth of real office rents has been identified, stepwise regression will be employed. Within the

95% confidence level, all of the variables will be put into the equation (6) in order to find the best

fit for the office real rental value movement in each South-east Asian city.

∆1 Rent it = A +∆1 Demand jit +∆1 Supply Iit + e (6)

Equations (4), (5) and (6) are the starting points for the empirical investigation.

Data: The availability of consistent time-series data on office rental values across South-east Asian

cities is limited; few series contain the necessary comprehensiveness and consistency across

markets and insightful analysis of the trends. Equally lacking are technically sound comparative

analyses. The data for office rental values of Singapore, Hong Kong, Kuala Lumpur and

Bangkok are obtained from Jones Lang LaSalle (Singapore JLL) series for investment-grade

office buildings in CBD areas, which is probably one of the very few consistent South-east

Asian data sets. They provide annual office rental value for the cities from 1988 to 2001, with

the exception of Taipei. The quarterly data are available only from 1995. Taipei office rental data

are from CB Richard Ellis (Taipei office) for prime office rental in the CBD area. Those rental

data refer to prime office rents of investment grade in the best locations. This type of office

building may be less sensitive to the economic cycle. Other factors might be more important to

16

its rental value, such as quality and location (Giussani et al. 1992). The data for changes in

office floor space area are also from Jones Lang LaSalle for the markets of Singapore, Hong

Kong, Bangkok and Kuala Lumpur. Taipei office floor space data are from CB Richard Ellis

(CBRE).

The data for consumer index data and real GDP for all of the countries are obtained by the local

government statistics departments, and are offered by Grosvenor Property, London.

Unemployment rates for five cities are obtained from Asian Development Bank Publications.

Prime lending rates and discount rates are used as interest rates in those five cities, obtained

from their central banks (Taiwan, Thailand, and Malaysia) and government departments of

statistics (Hong Kong and Singapore). Real GDP contributed by the finance sector is used as

the variable of service sector output, are obtained from Asian Development Bank Publications.

Empirical Results: The starting point for this analysis is to estimate Equation (4). Using office rental value in City i

regressed individually on 7 different variables (variable i), Table 3 exhibits the estimated

coefficient and the R2 statistic for each of 7 equations for each city in the analysis.

Table (3): OLS estimation of Rent it = A + B * Variables jit

City B R Square B R Square B R Square B R Square B R Square B R Square B R Square

Singapore

Hong Kong0.02059

(1.881)0.243

Taipei-0.005274

(1.938)0.273

169.657

(3.532)0.508

10.007788

(1.727)0.272

Kuala

Lumpur

0.008990

(2.421)0.328

-93.839

(5.493)0.715

70.841

(2.325)0.311

775.544

(2.332)0.312

0.03189

(3.114)0.447

Bangkok37.725

(2.907)0.413

-0.02007

(2.012)0.31

Service outputConsumer IndexF.S GDP Unemployment rates Interes t rates Lending Rates

There are several interesting points in Table (3): Firstly, there are few significant coefficients

estimates in Table (3) and there are no single significant variables in Singapore office

equations.

Secondly, service sector output seems significant in 4 out of 5 cities, except Singapore. Looking

at the significant variables, changes in floor space, which represents a supply-side variable,

only appear to be significant in the Taipei office market.

Thirdly, looking at the coefficient of determination (R2), a couple of the R2 exceed 40%,

17

especially in the unemployment rates equation for Taipei and Kuala Lumpur, which are

statistically satisfying results.

Fourthly, the similarity of the estimates for every city in unemployment rates and service sector

output is worth noticing. The estimated values in unemployment rates range between -93.839

and 169.657; and -0.02007 to 0.03189 for service sector output.

Finally, there are at least two significant variables in Taipei, Kuala Lumpur and Bangkok office

markets, especially in Kuala Lumpur markets.

A close examination of Table (3) reveals a number of other interesting results. Service sector

output appears to be the most common and important variable. It can be used to explain the

office rental value in Hong Kong, Taipei, Kuala Lumpur and Bangkok. In Taipei and Hong Kong,

service sector output can explain office rental value for 27.2% and 24.3% respectively. In Kuala

Lumpur and Bangkok, it can even explain 44.7% and 31% of office rents respectively. Floor

space only appears significant in the Taipei office market with the correct sign. Unemployment

rate shows a very strong explanatory power in Taipei and Kuala Lumpur markets, and their R2

are 50% and 71.5% respectively. There are 5 significant variables in the Kuala Lumpur office

market�s equations. They are real GDP, unemployment rates, prime lending rates, consumer

index and service sector output. Real GDP, prime lending rates and consumer index only

appear to be significant in the Kuala Lumpur office market. This result suggests there are high

demand-side drivers in the Kuala Lumpur office market. Interest rates and service

sectorsoutput have shown significant effects on office rents in Bangkok.

Largely, demand-side variables are significant in the case of 5 cites, according to Table (3),

apart from floor space in Taipei. This result shows that those demand-side variables which

influence South-east Asian cities rental markets differ from city to city, but service sector output

seems to be commonly significant.

For Hong Kong, Taipei, Kuala Lumpur and Bangkok, service sector output is statistically

important, although rents in Taipei and Kuala Lumpur can also be explained by unemployment

rates. The latter seems to be more important than service sector output in Taipei and Kuala

Lumpur office markets. While service sector output seems to predominate in the Hong Kong

office market and interest rates are the main explanatory factor in the Bangkok office market, at

R2 41.3%.

All the significant coefficients for floor space, real GDP, interest rates, prime lending rates and

consumer price index are correctly signed. Yet the unemployment rate in Taipei and the service

18

sector output in Bangkok show incorrect signs.

Time-series data are characteristics of trend. Table (3) might cause spurious correlation results

between variables. Thus, the first difference operation will transfer those variables in order to

eliminate the time series effect.

Previous analysis (Equation 4) has focused on establishing a pattern to the relationship

between office rental value and fluctuations in economic activity across the 5 selected cities,

without looking at the highly trended time-series data. To allow for spurious correlations, and to

model dynamic behaviour of office rents, Equation (4) was estimated once again, but using first

difference operation, which is in Equation (5). Again, this step can eliminate the effect of

time-series.

Table (4): OLS estimation of ∆1 Rent it = A +∆1 Variables jit

City B R Square B R Square B R Square B R Square B R Square B R Square B R Squa

Singapore-0.003471

(2.525)0.367

60.3295

(2.43)0.349

Hong Kong-0.01969

(2.54)0.364

Taipei107.442

(2.459)0.355

Kuala

Lumpur

43.456

(2.245)0.314

Bangkok na

Service outputF .S GDP Unemployment rates Interes t rates Lending Rates Consumer Index

re

Table (4) shows the result of equation (5). It demonstrates the dynamics of office rents and the

dynamics of macroeconomic variables in those 5 cities. Obviously, the number of statistically

significant variables in Table (4) is far less than in Table (3). Unlike in Table (3), Bangkok does

not have any significant variables to explain the variations of office rental value in Table (4).

Furthermore, changes in service sector output do not appear significant in any of those 5 cities.

Changes in floor space and changes in prime lending rates seem to be the most important

determinants of rent behaviour in 4 out of 5 cities.

The R2 statistics are also smaller than the value in Table (3). This suggests that those variables,

taken individually, account for between 31.4% and 36.7% of observed variation in changes of

South-east Asian office rental values. In Table (4), prime lending rates seem to be the main

variable in 3 out of 5 cities (Singapore, Taipei and Kuala Lumpur). In prime lending rates, all of

the coefficients are correctly signed. The magnitude of the estimated coefficient of prime

lending rates suggests that a 1% increase in prime lending rate over a year will increase rental

value around S$ 60 in Singapore, NT$ 107 in Taipei and Ringgit 43.456 in Kuala Lumpur.

Changes in floor space are significant in Singapore and Hong Kong office markets, which are

19

both correctly signed. This shows that when office floor space increases, then office rents fall,

and vice versa.

In equation (6), a high degree of correlation between changes in real GDP, unemployment rates,

interest rates and prime lending rates suggest that multicollinearity is the most likely outcome

from their simultaneous inclusion. However, the relatively short span of time-series data

available for this analysis made it necessary to include all those variables in the search

specification.

Table (5): OLS estimation of ∆1 Rent it = A +∆1 Demand jit +∆1 Supply Iit

R-Square

City

Singapore 0.367

Hong Kong

0.44

Taipei 0.821

Kuala Lumpur

0.314

Bangkok

43.456 (2.245)

B

1086.806 (2.802)

B B B B

140.234 (3.106)

B

NA

B

-0.00347 (2.52)

-0.00298 (5.74)

F.S GDP Unemployment rates

Interest rates Lending Rates

Consumer Index

Serv ice output

Table (5) exhibits annual changes in office rental values across South-east Asian cities. All

variables are included in the equation. Stepwise regression analysis is employed to find the

significant variables in each city. Table (5) shows that changes in floor space and prime lending

rates are significant variables in those cities. Changes in unemployment rates are only

significant in the Taipei office market. Yet again, there are no significant variables in the

Bangkok office market.

In Singapore, changes in floor space are significant (at 36.7%) in explaining office rental values

with the correct sign. In Hong Kong, prime lending rate is the only significant factor. When the

prime lending rate increases 1% per year, office rental value will increase HK$ 1086.8 per year.

The value of R2 in this equation is 44%. In Taipei, changes in floor space and unemployment

rates are significant. It is the only city which has more than one explanatory variable. The value

of adjusted R2 is 84%. Again, unemployment rate is wrongly signed. This might be because the

prime office market in Taipei is located in the prime CBD area, but unemployment rates are from

across the nation. The rental value has been very stable in the past 10 years and

unemployment rates are also very low in Taiwan. Taipei prime office markets might be less

sensitive to this wide proxy variable or even show a positive correlation. In Kuala Lumpur, prime

lending rate is the significant variable. An increase of 1% in the prime lending rate each year will

20

lead the office rental value upward for Ringgit 43.45. It can also explain 31% of the variation in

rents.

In Table (5), changes in floor space and changes in prime lending rates are the two main factors

explaining office rental markets in South-east Asian cities. These two variables are strongly

linked with development markets. In the past decade, changes in floor space (supply-side data)

have affected the regional office market (Chin 1999). Oversupply tends to be the common

problem across the region, which causes fluctuation in the rental level. In the late 1980s,

Singapore rental value fell sharply, because of oversupply. Table (5) results provide the

evidence of this. Prime lending rate is the key factor in financing in development. It seems to be

the main factor affecting the Hong Kong and Kuala Lumpur office markets. In the financial crisis

of 1997 across the region, oversupply and easy funding are the two main reasons which caused

the office markets to collapse. This can be seen from the empirical investigation in Table (5).

Conclusion: This paper has attempted to model the relationship between macro economic activities and

office rental movement in South-east Asian cities over the period 1988 to 2001. Office rental

value has been tested against some indicators of economic activity which have been used in

previous empirical studies. Because of significant data restrictions, the model tested the

influence of six demand-side variables and one supply-side variable. Those variables are Real

GDP, interest rates, prime lending rates, consumer price index, service service output,

unemployment rates and changes in office floor space.

Despite the short time span covered by the data, and deficiencies in its quality , there are still

some interesting findings. Changes in floor space and prime lending rates are key factors

determining rental values within the selected cities (apart from Bangkok). These two significant

factors are closely linked with development markets. This may partly be because of the

shortage of supply of floor space in the early 1990s, which drove rental value growth. Moreover,

easy availability of funding stimulates development schemes. As a consequence, rental value

dropped in 1997, because of the over-supply problem in South-east Asian cities. Compared to

European studies where real GDP, service sector output, and service sector unemployment

rates are the main determinants explaining office rental movements, none of those variables

show as significant in any of the five South-east Asian cities, apart from unemployment rates in

the Taipei office market (however, it is with a wrong sign). This might be because of the stability

of unemployment rates in the past 10 years in Taiwan, and the data used are at the national

level. The Bangkok office market does not have any significant variables to explain rental value

21

movement. This can possibly be explained by poor quality data or an immature market or both.

Bangkok and Kuala Lumpur are both regarded as emergent markets (Chin 2001). As a

consequence, those markets cannot be fully modelled by quantitative methods.

This research was hampered by the lack of consistent and comparable information for a

reasonable sample size, and short periods of time coverage. This was the case for both office

markets and economic data in South-east Asia, especially for the quality of office market data.

This problem can be solved by improving data quality in the future. As soon as there is a more

consistent, comparable and longer period of data, more sophisticated analysis can be

undertaken.

Office markets in South-east Asia have been opening up since the late 1980s, but this research

is still a relevant exercise, as it attempts to set up a benchmark for modelling South-east Asian

office markets. The analysis presented here clearly indicates a number of potential avenues for

future research. Firstly, the use of regional demand and supply indicators might improve the

robustness of the models. Secondly, the use of quarterly data could give for a better

understanding of the relative speed of the impact of changes in the key influences on market

dynamics. Thirdly, more sophisticated methods can be employed once the data quality and

range has improved.

The empirical results show that the explanatory power of the model is not very high. This may

partly be explained by the fact that some determinants of office rental values in these

South-east Asian cities cannot be quantified, and so cannot be included in the model.

Therefore, the next step will be to employ qualitative methods to examine the impact of

institutional factors on rental values across these cities.

Acknowledgements: The author would like to thank Professor Anthony Lavers, Professor Paul McNamara and Mr

Peter Dent for their guidance. Dr Richard Barkham and Mr Darren Rawcliffe from Grosvenor

Estates for their help in collecting much of the economic data employed in this paper, and Mr

Tony McGough in the City University who offered to share his extensive experience of

modelling office rental markets across Europe and valuable comments.

References:

22

1. Armitage L. (1996), International Property Investment Commercial Property Market

Behaviour in South-east Asia, paper presented in AREUEA Conference, Orlando, USA.

2. Asian Development Bank (2001), Key Indicators of Developing Asian and Pacific

Countries, Volume XXXII, Oxford University Press, Hong Kong.

3. Ball M., Lizieri C. and MacGregor B. D. (1998), The Economics of Commercial Property

Markets, Routledge, London.

4. Barras R. (1983), `A Simple Theoretical Model of the Office-Development Cycle`,

Environment and Planning A, 15, pp. 1381-1394.

5. Barras R. (1984), `The Office Development Cycle in London`, Land Development Studies,

1, pp 35-50.

6. Barras R. and Ferguson D. (1985), `A Spectral Analysis of Building Cycles in Britain`,

Environment and Planning A, 17, pp. 1369-1391.

7. Barras R. and Ferguson D. (1987), `Dynamic Modelling of the Building Cycle: 1.

Theoretical Framework`, Environment and Planning A, 19, pp. 353-367.

8. Barras R. and Ferguson D. (1987), ̀ Dynamic Modelling of the Building Cycle: 2. Empirical

Results`, Environment and Planning A, 19, pp. 493-520.

9. Barras R. (1994), `Property and the Economic Cycles: Building Cycles Revisited`, Journal

of Property Research, 11(4), pp. 183-197.

10. BisChoff C, (1970), “A Model of Non-residential Construction in the United States”

American Economic Review, Papers and Proceedings, 60(2), pp. 10-17.

11. Blake N, Matysiak G. and Lizieri C. (2000), Research Report: Forecasting Office Supply

and Demand, R.I.C.S., London.

12. Boyd T (1995), 'The Concept and Determination of Investment Value', Paper Presented at

the First Pacific Rim Real Estate Society Conference, Melbourne Australia.

13. Brown G. R and Matysiak G. A. (2000), Real Estate Investment: A Capital market

Approach, Financial Times Prentice Hall, London.

14. Burns A. F. and Mitchell W. C. (1946), Measuring Business Cycles, National Bureau of

Economic Research, New York.

15. Chang C. and Huang M. (1998), A Study of Office Rent in Taipei, Paper Presented at

Chinese Society of Housing 8th Annual Conference, Taipei Taiwan. (In Chinese)

16. Chin W. (1999), Dissertation: The Assessment of Country Risk: The Case of Southeast

Asia, Department of Land Management and Development, University of Reading, Reading

17. Chin W. (2001), �Factors for Determining the Office Markets in Southeast Asian Cities with

reference to Hong Kong, Singapore, Taipei, Bangkok and Kuala Lumpur”, Paper Presented

at Cutting Edge Conference, Oxford, U.K.

18. Chin W. (2002), �Determinants for Office Investment Markets in Southeast Asian Cities,

with reference to Hong Kong, Singapore, Taipei, Bangkok and Kuala Lumpur’, Paper

Presented at Pacific Rims Real Estate Society Conference, Christchurch, New Zealand.

23

19. Clapp J. M. (1993), Dynamics of Office Markets: Empirical Findings and Research issues,

The Urban Institute Press, Washington.

20. Cushman & Wakefield Research, Asia Overview at http://www.cushwakeasia.com.

21. D'Arcy E. (1998), Lecture Note: Economic for Planning and Land Management, The

University of Reading, Reading.

22. D’Arcy E. (1998), Lecture Note: International Real Estate Markets, The University of

Reading, Reading.

23. D'Arcy E., McGough T. and Tsolocos S. (1994), Modelling the Determinants of Office

Rental Values in Major European Cities , Discussion papers in Urban and Regional

Economics, Series C, Vol. VII, No. 99, Department of Economics, The University of

Reading, Reading.

24. D'Arcy E., McGough T. and Tsolocos S. (1997), `National Economic Trends, Market Size

and City Growth Effects on European Office Rents`, Journal of Property Research, 13 (4),

pp 297-308.

25. D'Arcy E., McGough T. and Tsolocos S. (1998), An Econometric Analysis and Forecast

of the Office Rental Cycle in the Dublin Area, Discussion papers in Urban and Regional

Economics, Series C, Vol. XI, No. 33, Department of Economics, The University of Reading,

Reading.

26. DiPasquale D. and Wheaton W. C. (1992), ̀ The Markets fir Real Estate Assets and Space:

A Conceptual Framework`, Journal of the American Real Estate and Urban Economics

Association, 22(1), pp. 181-197.

27. Dobson S. M. and Goddard J.A. (1992), ̀ The Determinants of Commercial Property Price

and Rents`, Bulletin of Economic Research, 44(4), pp. 301-321.

28. DTZ Research, Property Market Report, various issues.

29. Fisher J. D. (1992), 'Integrating Research on Markets for Space and Capital', American

Real Estate and Urban Economic Association Journal, 20(2), pp 161-180.

30. Gardliner C., Henneberry J. (1988), `The Development of a Simple Regional Model of

Office Rent Prediction`, Journal of Property Valuation and Investment, 7, pp 36-52.

31. Gardliner C., Henneberry J. (1991), 'Predicting Regional Office Rents Using

Habit-Persistence Theories', Journal of Property Valuation and Investment 9(3), pp 215 �

226.

32. Giussani B., Hsia M., Tsolocas S. (1993a), `A Comparative Analysis of the Major

Determinants of Office Rental values in Europe`, Journal of Property Valuation and

Investment, 11, pp 157-173.

33. Giussani B., Hsia M., Tsolocas S. (1993b), ‘ Economic Growth, uncertainty and

commercial property in Europe: A Comparative Analysis of the Major Determinants of

Office Rental Values’, Discussion Papers in Urban and Regional Economics, Series C, Vol

V, No. 77, The University of Reading.

24

34. Hendershott P. H. (1996), `Rental Adjustment and Valuation Overbuilt Markets: Evidence

from Sydney Office Market`, Journal of Urban Economics, 39, pp. 51-67.

35. Hendershott P. H., Lizieri C. M. and Matysiak G. A. (1996), `Modelling the London Office

Market`, Paper Presented at RICS Cutting Edge Conference, September 1996, Bristol.

36. Hekman J. (1985), 'Rental Price Adjustment and Investment in the Office Market',

American Real Estate and Urban Economics Association Journal, Vol. 13 (1), pp31 � 47.

37. Hetherington F. T. (1988), Forecasting of Rents, in MacLeary A.R. and Nathacumaran N.

(Eds), Property Investment Theory, Spon, London.

38. Higgins D. M. (1999), `Commercial Real Estate Cycles: A Conceptual Frame Work,

Measuring and Comparing Commercial Real Estate Cycles`, Australia Property Journal,

February, pp. 421- 427.

39. Higgins D. M. and Valence G. (2000), `Australian Business Cycles and Commercial

Property markets: Some Empirical Evidence Over Four Decades`, Pacific Rim Property

Research Journal, 6 (19), pp. 57-66.

40. Hillier Parker (2000), Global market Rents: Office Rents and Occupancy Costs Worldwide,

CB Hillier Parker, London.

41. Jones Lang Wootton (JLW) Research, Report Property Market, various issue.

42. Jones Lang Wootton (JLW) Research, Asia Pacific Property Digest, various issue.

43. Jones Lang LaSalle (JLL) Research, Report Property Market, various issue.

44. Kelly H. (1983), 'Forecasting Office Space Demand in Urban Areas', Real Estate Review,

Vol. 13, pp. 87-95.

45. Keogh G. (1994), 'Use and Investment Markets in British Real Estate', Journal of property

Valuation and Investment, 12 (4), pp. 58-72.

46. Keogh G., McGough T. and Tsolocas S. (1998), `Modelling Use, Investment, and

Development in the British Office Market`, Environment and Planning A, Vol. 30, pp

1409-1427.

47. Knight Frank Research, Asia Pacific Property Review, various issues.

48. Lee S. (1998), Lecture Note: Forecasting, The University of Reading, Reading.

49. MacFarlane J. (1998), `An Examination of Property Cycles in the Office Markets of

Selected Australian Capital Cities`, Paper Presented in Pacific Rim Real Estate Society

Conference 1998, Perth.

50. MacFarland J. and Moon S. (1999), `Modelling of Office markets in Australia`, Paper

Presented in Pacific Rim Real Estate Society Conference 1999, Kuala Lumpur.

51. MacFarland J. and Moon S. (2000), `Characteristics of Australian Commercial Property

markets: Towards an Improved Forecasting Model`, Paper Presented in Pacific Rim Real

Estate Society Conference 2000, Sydney.

52. McClure K. (1991), `Estimating Occupied Office Space: Comparing Alternative Forecast

Methodologies`, The Journal of Real Estate Research, 6(3), pp. 305-314.

25

53. McCue T. E. and Kling J. L. (1987), `Office Building Investment and the Macroeconomy:

Empirical Evidence 1973-1985`, Journal of the American Real Estate and Urban

Economics Association, 15(3), pp. 234-255.

54. McCue T. E. and Kling J. L. (1994), `Real Estate Returns and the Macroeconomy: Some

Empirical Evidence from Real Estate Investment Trust Data, 1972-1991`, The Journal of

Real Estate Research, 9(3), pp. 277-287.

55. McGough T. and Tsolocas S (1994), `Forecasting Office Rental Values Using Vector

Auto-regressive Models`, The Cutting Edge 1994, Proceedings of the Cutting Edge

Property Research of RICS, The Royal Institute of Chartered Surveyors, London, pp.

303-319.

56. McGough T. and Tsolocas S. (1994), `Forecasting Commercial Rental Values Using

ARIMA Models`, Journal of Property Valuation and Investment, 13(5), pp. 6-22.

57. McGough T. and Tsolocas S. (1995), `Property Cycles in the UK: an Empirical

Investigation of the Stylised Facts`, Journal of Property Finance, 6(4), pp 45-62.

58. McGough T. and Tsolocas S. (1997), `The Stylised Facts of the UK Commercial Building

Cycles`, Environment and Planning A, 29, pp. 485-500.

59. McGough T. and Tsolocas S. (1999), `Rational Expectations, Uncertainty and Cyclical

Activity in British Office Market`, Urban Studies, Vol. 36, No. 7, pp. 1137-1149.

60. McGough T., Olkkonen O. and Tsolocas S. (1998), The Cyclical Behaviour of Office

Rents in Helsinki, European Real Estate Society's Annual Conference, Maastricht, The

Netherlands.

61. Morrison N. (1994), `Forecasting Local Property Market in Kent`, The Cutting Edge 1994,

Proceedings of the Cutting Edge Property Research of RICS, The Royal Institute of

Chartered Surveyors, London, pp. 271-289.

62. Newell G. and Chau K.W. (2000), Using Style Analysis to Assess Hong Kong Property

Performance, Paper Presented at Asian Real Estate Society Fifth Annual Conference,

Beijing China.

63. Parker D., MacFarlane J. and Whiley (2001), Forecasting Property Market Cycles: An

Application of the RICS Model to Australian Property Markets, Paper Presented at RICS

Cutting Edge Conference, Oxford, U.K..

64. Pollakowski H. O. (1992), `Research on Office Markets`, Journal of American Real Estate

and Urban Economics Association, 20(1), pp. 155-160.

65. Pollakowski H. O., Wachter S. M. and Lynford L. (1992), `Did Office Market Size Matter

is the 1980s? A Time Series Cross-Sectional Analysis of Metropolitan Area Office Markets`,

Journal of American Real Estate and Urban Economics Association, 20(1), pp. 303-324.

66. Rosen K. T. (1984), `Toward A Model of The Office Building Sector`, Journal of the

American Real Estate and Urban Economics Association, 12 (3), pp. 261-269.

67. Richard Ellis, Market Index, various issues.

26

68. RICS (1994), Understanding The Property Cycle: Economic Cycles and Property Cycles,

RICS, London.

69. RICS and Society of Property Researchers (1989), The Application of Forecasting

Techniques to the Property Market, RICS and Society of Property Researchers, London.

70. Tsolocas S. (1998), Lecture Notes: International Real Estate Markets, The University of

Reading, Reading.

71. Walker A and Chung L. W. (1990), Hong Kong: Property, Construction and the Economy,

RICS, London.

72. Walker A. and Flanagan R. (1991), Property and Construction in Asia Pacific: Hong Kong,

Japan, Singapore, Blackwell, Oxford.

73. Wheaton W. C. (1987), ̀ The Cyclic Behaviour of the National Office Market`, Journal of the

American Real Estate and Urban Economics Association, 15(4), pp. 281-299.

74. Wheaton W. C. and Torto R. G. (1988), `Vacancy Rates and the Future of Office Rents`,

Journal of the American Real Estate and Urban Economics Association, 16(4), pp.

430-436.

75. Wheaton W. C. and Torto R. G. (1990), `An Investment Model of the Demand and Supply

for Industrial Real Estate `, Journal of the American Real Estate and Urban Economics

Association, 18(4), pp. 530-547.

76. Wheaton W. C., Torto R. G. and Evans P. (1997), `The Cyclic Behaviour of the Greater

London Office Market`, Journal of Real Estate Finance and Economics, 15 (19), pp. 77-92.

77. Wong R. (2002), �A New Dimension in Property Forecasting: Conceptual Framework for an

Integrated Modular Approach”, Paper Presented at PRRES Conference, Christchurch,

New Zealand.

78. Yong S. (2000), Impact of the Asian Crisis Including the Malaysian Property Market

Perspective, Pacific-Rim Real Estate Society Conference, Sydney.

27

Appendices: Appendix 1: Singapore Investment Grade Office Rental Market

200

400

600

800

1000

1200

88 89 90 91 92 93 94 95 96 97 98 99 00 01

RENT

Year

S$ per square meter/per year

Source: Jones Lang LaSalle

28

Appendix 2: Hong Kong Investment Grade Office Rental Market

2000

4000

6000

8000

10000

12000

88 89 90 91 92 93 94 95 96 97 98 99 00 01

RENT

Year

HK$ per square meter/per year

Source: Jones Lang LaSalle

29

Appendix 3: Taipei Prime Office Rental Market

2200

2400

2600

2800

3000

88 89 90 91 92 93 94 95 96 97 98 99 00 01

RENT

Year

NT$ Per Ping/ Per Month

Source: CB Richard Ellis

30

Appendix 4: Kuala Lumpur Investment Grade Office Rental Market

100

200

300

400

500

600

88 89 90 91 92 93 94 95 96 97 98 99 00 01

RENT

Year

RM Per Square Meter/ Per Year

Source: Jones Lang LaSalle

31

Appendix 5: Bangkok Investment Grade Office Rental Market

1000

2000

3000

4000

5000

6000

7000

8000

88 89 90 91 92 93 94 95 96 97 98 99 00 01

RENT

Year

Baht Per Square Meter/ Per Year

Source: Jones Lang LaSalle